Durham police to use AI for custody decisions
System has an 89 per cent success rate in identifying suspects who are likely to offend

Police in Durham are preparing to use artificial intelligence (AI) to assist officers deciding whether or not to send a suspect into custody, reports the BBC.
A system has been developed to categorise suspects into "low, medium or high risk of offending". It has been developed using five years of criminal history data.
Sheena Urwin, head of criminal justice at Durham Constabulary, told the BBC: "I imagine in the next two to three months we'll probably make it a live tool to support officers' decision making".
Subscribe to The Week
Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

Sign up for The Week's Free Newsletters
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
Police trialled the harm assessment risk tool (Hart) for a two-year period starting in 2013, says Alphr, during which researchers discovered it had a 98 per cent success rate in identifying low-risk subjects and an 89 per cent rate for high-risk subjects.
It's decisions are based on factors such as "seriousness of alleged crime and previous criminal history".
Hart "leans towards a cautious outlook", says Alphr, so it is more likely to label a suspect as medium or high-risk, reducing the danger of "releasing dangerous criminals".
Such technology is becoming a vital tool in helping police in their investigations.
Last month, a man was charged with murdering his wife after investigators were able to work out her final moments using her Fitbit health tracker.
Information on how many steps the victim had walked indicated she had been active for an hour after the time her husband said she died, says The Guardian.
It also suggested she had "traveled more than 1,200ft after arriving home", adds the paper, while her husband said she was murdered by intruders immediately after arriving.
Sign up for Today's Best Articles in your inbox
A free daily email with the biggest news stories of the day – and the best features from TheWeek.com
-
Fake AI job seekers are flooding U.S. companies
In the Spotlight It's getting harder for hiring managers to screen out bogus AI-generated applicants
By Theara Coleman, The Week US
-
How might AI chatbots replace mental health therapists?
Today's Big Question Clients form 'strong relationships' with tech
By Joel Mathis, The Week US
-
What are AI hallucinations?
The Explainer Artificial intelligence is known for making things up – and that can cause real damage
By Elizabeth Carr-Ellis, The Week UK
-
The backlash against ChatGPT's Studio Ghibli filter
The Explainer The studio's charming style has become part of a nebulous social media trend
By Theara Coleman, The Week US
-
Not there yet: The frustrations of the pocket AI
Feature Apple rushes to roll out its ‘Apple Intelligence’ features but fails to deliver on promises
By The Week US
-
OpenAI's new model is 'really good' at creative writing
Under the Radar CEO Sam Altman says he is impressed. But is this merely an attempt to sell more subscriptions?
By Theara Coleman, The Week US
-
Could artificial superintelligence spell the end of humanity?
Talking Points Growing technology is causing growing concern
By Devika Rao, The Week US
-
Space-age living: The race for robot servants
Feature Meta and Apple compete to bring humanoid robots to market
By The Week US